The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands like docker ps, npx supabase status, and grep to verify the state of local services and analyze project source code for configuration patterns. These are standard operations for a local debugging tool.
[PROMPT_INJECTION]: The skill provides a workflow for auditing localStorage usage that scans the local src/ directory. This creates a surface for indirect prompt injection if application files contain malicious instructions intended for the agent. 1. Ingestion points: The grep command in SKILL.md (Workflow 8) reads content from the src/ directory. 2. Boundary markers: Absent. 3. Capability inventory: The skill includes workflows that use psql for database modification, curl for network requests, and bash for command execution. 4. Sanitization: No sanitization is performed on the ingested source code content.
[SAFE]: The SQL and JavaScript snippets provided are standard administrative patterns for Supabase monitoring and RLS validation. The skill uses placeholders for sensitive values and does not perform any unauthorized data exfiltration or persistence operations.

supabase-debugger